Document 10821924

advertisement
Hindawi Publishing Corporation
Abstract and Applied Analysis
Volume 2011, Article ID 251612, 19 pages
doi:10.1155/2011/251612
Research Article
Strong Convergence Theorems for Families of
Weak Relatively Nonexpansive Mappings
Yekini Shehu
Mathematics Institute, African University of Science and Technology, Abuja, Nigeria
Correspondence should be addressed to Yekini Shehu, deltanougt2006@yahoo.com
Received 30 September 2010; Revised 25 November 2010; Accepted 15 February 2011
Academic Editor: Jean Pierre Gossez
Copyright q 2011 Yekini Shehu. This is an open access article distributed under the Creative
Commons Attribution License, which permits unrestricted use, distribution, and reproduction in
any medium, provided the original work is properly cited.
We construct a new Halpern type iterative scheme by hybrid methods and prove strong
convergence theorem for approximation of a common fixed point of two countable families of
weak relatively nonexpansive mappings in a uniformly convex and uniformly smooth real Banach
space using the properties of generalized f-projection operator. Using this result, we discuss strong
convergence theorem concerning general H-monotone mappings. Our results extend many known
recent results in the literature.
1. Introduction
Let E be a real Banach space with dual E∗ and let C be nonempty, closed and convex subset
of E. A mapping T : C → C is called nonexpansive if
T x − T y ≤ x − y,
∀x, y ∈ C.
1.1
A point x ∈ C is called a fixed point of T if T x x. The set of fixed points of T is denoted by
FT : {x ∈ C : T x x}.
∗
We denote by J the normalized duality mapping from E to 2E defined by
2 Jx f ∈ E∗ : x, f x2 f .
1.2
The following properties of J are well known. The reader can consult 1–3 for more details.
1 If E is uniformly smooth, then J is norm-to-norm uniformly continuous on each
bounded subset of E.
2
Abstract and Applied Analysis
2 Jx /
∅, x ∈ E.
3 If E is reflexive, then J is a mapping from E onto E∗ .
4 If E is smooth, then J is single valued.
Throughout this paper, we denote by φ, the functional on E × E defined by
2
φ x, y x2 − 2 x, J y y ,
∀x, y ∈ E.
1.3
From 4, in uniformly convex and uniformly smooth Banach spaces, we have
2
2 2
≤ φ x, y ≤ x y ,
x − y
∀x, y ∈ E.
1.4
Definition 1.1. Let C be a nonempty subset of E and let {Tn }∞
n0 be a countable family of
mappings from C into E. A point p ∈ C is said to be a asymptotic fixed point of {Tn }∞
n0 if
∞
C contains a sequence {xn }n0 which converges weakly to p and limn → ∞ xn − Tn xn 0.
∞
n }∞
The set of asymptotic fixed points of T is denoted by F{T
n0 . We say that {Tn }n0 is
countable family of relatively nonexpansive mappings see, e.g., 5 if the following conditions
are satisfied:
R1 F{Tn }∞
/ ∅;
n0 R2 φp, Tn x ≤ φp, x, for all x ∈ C, p ∈ FTn , n ≥ 0;
∞
R3 ∞
n0 FTn F{Tn }n0 .
Definition 1.2. A point p ∈ C is said to be a strong asymptotic fixed point of {Tn }∞
n0 if C contains
a sequence {xn }∞
n0 which converges strongly to p and limn → ∞ xn − Tn xn 0. The set of
∞
n }∞
strong asymptotic fixed points of T is denoted by F{T
n0 . We say that a mapping {Tn }n0
is countable family of weak relatively nonexpansive mappings see, e.g., 5 if the following
conditions are satisfied:
R1 F{Tn }∞
/ ∅;
n0 R2 φp, Tn x ≤ φp, x, for all x ∈ C, p ∈ FTn , n ≥ 0;
∞
R3 ∞
n0 FTn F{Tn }n0 .
Definition 1.3. Let C be a nonempty subset of E and let T be a mapping from C into E. A
point p ∈ C is said to be an asymptotic fixed point of T if C contains a sequence {xn }∞
n0 which
converges weakly to p and limn → ∞ xn − T xn 0. The set of asymptotic fixed points of T is
. We say that a mapping T is relatively nonexpansive see, e.g., 6–11 if the
denoted by FT
following conditions are satisfied:
R1 FT /
∅;
R2 φp, T x ≤ φp, x, for all x ∈ C, p ∈ FT ;
.
R3 FT FT
Definition 1.4. A point p ∈ C is said to be an strong asymptotic fixed point of T if C contains a
sequence {xn }∞
n0 which converges strongly to p and limn → ∞ xn − T xn 0. The set of strong
Abstract and Applied Analysis
3
. We say that a mapping T is weak relatively
asymptotic fixed points of T is denoted by FT
nonexpansive see, e.g., 12, 13 if the following conditions are satisfied:
R1 FT /
∅;
R2 φp, T x ≤ φp, x, for all x ∈ C, p ∈ FT ;
.
R3 FT FT
Definition 1.3 Definition 1.4, resp. is a special form of Definition 1.1 Definition 1.2,
resp. as Tn ≡ T , for all n ≥ 0. Furthermore, Su et al. 5 gave an example which is a
countable family of weak relatively nonexpansive mappings but not a countable family of
relatively nonexpansive mappings. It is obvious that relatively nonexpansive mapping is
weak relatively nonexpansive mapping. In fact, for any mapping T : C → C, we have FT ⊂
⊂ FT
. Therefore, if T is relatively nonexpansive mapping, then FT FT
FT
.
FT
Kang et al. 12 gave an example of a weak relatively nonexpansive mapping which is not
relatively nonexpansive.
In 9, Matsushita and Takahashi introduced a hybrid iterative scheme for approximation of fixed points of relatively nonexpansive mapping in a uniformly convex real Banach
space which is also uniformly smooth: x0 ∈ C,
yn J −1 αn Jxn 1 − αn JT xn ,
Hn w ∈ C : φ w, yn ≤ φw, xn ,
1.5
Wn {w ∈ C : xn − w, Jx0 − Jxn ≥ 0},
xn1 ΠHn ∩Wn x0 ,
n ≥ 0.
They proved that {xn }∞
∅.
n0 converges strongly to ΠFT x0 , where FT /
In 14, Plubtieng and Ungchittrakool introduced the following hybrid projection
algorithm for a pair of relatively nonexpansive mappings: x0 ∈ C,
1
2
3
zn J −1 βn Jxn βn JT xn βn JSxn ,
yn J −1 αn Jx0 1 − αn Jzn ,
Cn z ∈ C : φ z, yn ≤ φz, xn αn x0 2 2w, Jxn − Jx0 ,
1.6
Qn {z ∈ C : xn − z, Jxn − Jx0 ≤ 0},
xn1 PCn ∩Qn x0 ,
1
2
3
1
2
3
where {αn }, {βn }, {βn } and {βn } are sequences in 0, 1 satisfying βn βn βn 1 and T
and S are relatively nonexpansive mappings and J is the single-valued duality mapping on
E. They proved under the appropriate conditions on the parameters that the sequence {xn }
generated by 1.6 converges strongly to a common fixed point of T and S.
Recently, Li et al. 15 introduced the following hybrid iterative scheme for
approximation of fixed points of a relatively nonexpansive mapping using the properties
4
Abstract and Applied Analysis
of generalized f-projection operator in a uniformly smooth real Banach space which is also
uniformly convex: x0 ∈ C,
Cn1
yn J −1 αn Jxn 1 − αn JT xn ,
w ∈ Cn : G w, Jyn ≤ Gw, Jxn ,
f
xn1 ΠCn1 x0 ,
1.7
n ≥ 1.
They proved a strong convergence theorem for finding an element in the fixed points set of
T . We remark here that the results of Li et al. 15 extended and improved on the results of
Matsushita and Takahashi, 9.
Quite recently, motivated by the results of Matsushita and Takahashi 9 and Plubtieng
and Ungchittrakool 14, Su et al. 5 proved the following strong convergence theorem
by Halpern type hybrid iterative scheme for approximation of common fixed point of two
countable families of weak relatively nonexpansive mappings in uniformly convex and
uniformly smooth Banach space.
Theorem 1.5. Let E be a uniformly convex real Banach space which is also uniformly smooth. Let C
∞
are two countable families
be a nonempty, closed and convex subset of E. Suppose {Tn }∞
n1 and {Sn }n1 ∞
of weak relatively nonexpansive mappings of C into itself such that F : ∞
n1 FSn n1 FTn ∩
∞
∅. Suppose {xn }∞
n1 FSn /
n0 is iteratively generated by x0 ∈ C,
1
2
3
zn J −1 βn Jx0 βn JTn xn βn JSn xn ,
yn J −1 αn Jzn 1 − αn Jxn ,
1
1
Cn w ∈ Cn−1 ∩ Qn−1 : φ w, yn ≤ 1 − αn βn φw, xn αn βn φw, x0 ,
C0 w ∈ C : φ w, y0 ≤ φw, x0 ,
1.8
Qn {w ∈ Cn−1 ∩ Qn−1 : xn − w, Jx0 − Jxn ≥ 0},
Q0 C,
xn1 ΠCn ∩Qn x0 ,
n ≥ 1,
with the conditions
1
i limn → ∞ βn 0;
2 3
ii lim supn → ∞ βn βn > 0.
Then, {xn }∞
n0 converges strongly to ΠF x0 .
Motivated by the above mentioned results and the ongoing research, it is our purpose
in this paper to prove a strong convergence theorem by Halpern type iterative scheme for
two countable families of weak relatively nonexpansive mappings in a uniformly convex
and uniformly smooth real Banach space using the properties of generalized f-projection
operator. Our results extend the results of Su et al. 5 and many other recent known results
in the literature.
Abstract and Applied Analysis
5
2. Preliminaries
Let E be a real Banach space. The modulus of smoothness of E is the function ρE : 0, ∞ →
0, ∞ defined by
1 x y x − y − 1 : x ≤ 1, y ≤ t .
ρE t : sup
2
2.1
E is uniformly smooth if and only if
lim
t→0
ρE t
0.
t
2.2
Let dim E ≥ 2. The modulus of convexity of E is the function δE : 0, 2 → 0, 1 defined by
x y
: x y 1; x − y .
δE : inf 1 − 2 2.3
E is uniformly convex if for any ∈ 0, 2, there exists a δ δ > 0 such that if x, y ∈ E with
x ≤ 1, y ≤ 1 and x − y ≥ , then 1/2x y ≤ 1 − δ. Equivalently, E is uniformly
convex if and only if δE > 0 for all ∈ 0, 2. A normed space E is called strictly convex if
for all x, y ∈ E, x /
y, x y 1, we have λx 1 − λy < 1, for all λ ∈ 0, 1.
Let E be a smooth, strictly convex, and reflexive real Banach space and let C be a
nonempty, closed, and convex subset of E. Following Alber 16, the generalized projection
ΠC from E onto C is defined by
ΠC x : arg min φ y, x ,
y∈C
∀x ∈ E.
2.4
The existence and uniqueness of ΠC follows from the property of the functional φx, y and
strict monotonicity of the mapping J see, e.g., 3, 4, 16–18. If E is a Hilbert space, then ΠC
is the metric projection of H onto C.
Next, we recall the concept of generalized f-projector operator, together with its
properties. Let G : C × E∗ → R ∪ {∞} be a functional defined as follows:
2
G ξ, ϕ ξ2 − 2 ξ, ϕ ϕ 2ρfξ,
2.5
where ξ ∈ C, ϕ ∈ E∗ , ρ is a positive number and f : C → R ∪ {∞} is proper, convex,
and lower semicontinuous. From the definitions of G and f, it is easy to see the following
properties:
i Gξ, ϕ is convex and continuous with respect to ϕ when ξ is fixed;
ii Gξ, ϕ is convex and lower semicontinuous with respect to ξ when ϕ is fixed.
6
Abstract and Applied Analysis
Definition 2.1 Wu and Huang 19. Let E be a real Banach space with its dual E∗ . Let C be
f
a nonempty, closed, and convex subset of E. We say that ΠC : E∗ → 2C is a generalized
f-projection operator if
f
ΠC ϕ
u ∈ C : G u, ϕ inf G ξ, ϕ ,
ξ∈C
∀ϕ ∈ E∗ .
2.6
For the generalized f-projection operator, Wu and Huang 19 proved the following theorem
basic properties.
Lemma 2.2 Wu and Huang 19. Let E be a real reflexive Banach space with its dual E∗ . Let C be
a nonempty, closed, and convex subset of E. Then the following statements hold:
f
i ΠC is a nonempty closed convex subset of C for all ϕ ∈ E∗ ;
f
ii if E is smooth, then for all ϕ ∈ E∗ , x ∈ ΠC if and only if
x − y, ϕ − Jx ρf y − ρfx ≥ 0,
∀y ∈ C;
2.7
iii if E is strictly convex and f : C → R ∪ {∞} is positive homogeneous (i.e., ftx tfx
f
for all t > 0 such that tx ∈ C where x ∈ C), then ΠC is a single valued mapping.
Fan et al. 20 showed that the condition f is positive homogeneous which appeared
in Lemma 2.2 can be removed.
Lemma 2.3 Fan et al. 20. Let E be a real reflexive Banach space with its dual E∗ and C a
f
nonempty, closed and convex subset of E. Then if E is strictly convex, then ΠC is a single valued
mapping.
Recall that J is a single valued mapping when E is a smooth Banach space. There exists
a unique element ϕ ∈ E∗ such that ϕ Jx for each x ∈ E. This substitution in 4.3 gives
Gξ, Jx ξ2 − 2ξ, Jx x2 2ρfξ.
2.8
Now, we consider the second generalized f-projection operator in a Banach space.
Definition 2.4. Let E be a real Banach space and C a nonempty, closed and convex subset of E.
f
We say that ΠC : E → 2C is a generalized f-projection operator if
f
ΠC x
u ∈ C : Gu, Jx inf Gξ, Jx ,
ξ∈C
∀x ∈ E.
2.9
Abstract and Applied Analysis
7
Obviously, the definition of {Tn }∞
n0 is a countably family of weak relatively nonexpansive
mappings is equivalent to
R 1 F{Tn }∞
/ ∅;
n0 R 2 Gp, JTn x ≤ Gp, Jx, for all x ∈ C, p ∈ FTn , n ≥ 0.
∞
R 3 ∞
n0 FTn F{Tn }n0 .
Lemma 2.5 Li et al. 15. Let E be a Banach space and f : E → R ∪ {∞} be a lower
semicontinuous convex functional. Then there exists x∗ ∈ E∗ and α ∈ R such that
fx ≥ x, x∗ α,
∀x ∈ E.
2.10
f
We know that the following lemmas hold for operator ΠC .
Lemma 2.6 Li et al. 15. Let C be a nonempty, closed, and convex subset of a smooth and reflexive
Banach space E. Then the following statements hold:
i ΠC xf is a nonempty closed, and convex subset of C for all x ∈ E;
ii for all x ∈ E, x ∈ ΠC xf if and only if
x − y, Jx − J x ρf y − ρfx ≥ 0,
∀y ∈ C;
2.11
iii if E is strictly convex, then ΠC xf is a single valued mapping.
Lemma 2.7 Li et al. 15. Let C be a nonempty, closed, and convex subset of a smooth and reflexive
f
Banach space E. Let x ∈ E and x ∈ ΠC x. Then
φ y, x Gx,
Jx ≤ G y, Jx ,
∀y ∈ C.
2.12
Lemma 2.8 Su et al. 5. Let C be a nonempty, closed, and convex subset of a smooth, strictly
convex Banach space E. Let T be a weak relatively nonexpansive mapping of C into itself. Then FT is closed and convex.
Also, this following lemma will be used in the sequel.
Lemma 2.9 Kamimura and Takahashi 4. Let C be a nonempty, closed and convex subset of a
∞
smooth, uniformly convex Banach space E. Let {xn }∞
n1 and {yn }n1 be sequences in E such that either
∞
{xn }∞
n1 or {yn }n1 is bounded. If limn → ∞ φxn , yn 0, then limn → ∞ xn − yn 0.
Lemma 2.10 Cho et al. 21. Let E be a uniformly convex real Banach space. For arbitrary r > 0,
let Br 0 : {x ∈ E : x ≤ r} and λ, μ, γ ∈ 0, 1 such that λ μ γ 1. Then, there exists a
continuous strictly increasing convex function
g : 0, 2r −→ R,
g0 0
2.13
such that for every x, y, z ∈ Br 0, the following inequality holds:
λx μy γz2 ≤ λx2 μy2 − λμg x − y .
2.14
8
Abstract and Applied Analysis
For the rest of this paper, the sequence {xn }∞
n0 converges strongly to p will be denoted by
1 2 3
1
2
3
xn → p as n → ∞ and we will assume that βn , βn , βn ∈ 0, 1 such that βn βn βn 1, for all n ≥ 0.
3. Main Results
Theorem 3.1. Let E be a uniformly convex real Banach space which is also uniformly smooth. Let
∞
C be a nonempty, closed, and convex subset of E. Suppose {Tn }∞
n0 and {Sn }n0 aretwo countable
families of weak relatively nonexpansive mappings of C into itself such that F : ∞
n0 FTn ∩
FS
∅.
Let
f
:
E
→
R
be
a
convex
and
lower
semicontinuous
mapping
with C ⊂
∞
/
n
n0
is
iteratively
generated
by
x
∈
C,
C
C,
intDf and suppose {xn }∞
0
0
n0
1
2
3
zn J −1 βn Jx0 βn JTn xn βn JSn xn ,
Cn1
yn J −1 αn Jzn 1 − αn Jxn ,
1
1
w ∈ Cn : G w, Jyn ≤ 1 − αn βn Gw, Jxn αn βn Gw, Jx0 ,
f
xn1 ΠCn1 x0 ,
3.1
n ≥ 0,
with the conditions
1
i limn → ∞ βn 0;
2
3
ii 0 < b ≤ βn ≤ 1, 0 < d ≤ βn ≤ 1;
iii 0 < α ≤ αn ≤ 1 for some α ∈ 0, 1.
f
Then, {xn }∞
n0 converges strongly to ΠF x0 .
Proof. We first show that Cn , for all n ≥ 0 is closed and convex. It is obvious that C0 C is
closed and convex. Thus, we only need to show that Cn is closed and convex for each n ≥ 1.
1
1
Since Gw, Jyn ≤ 1 − αn βn Gw, Jxn αn βn Gw, Jx0 is equivalent to
2
1
1
1
1
≤ 1 − αn βn xn 2 αn βn x0 2 − yn . 3.2
2 w, 1 − αn βn Jxn αn βn Jx0 − Jyn
This implies that Cn1 is closed and convex for all n ≥ 0.
We next show that F ⊂ Cn , for all n ≥ 0. For n 0, we have F ⊂ C C0 . Then for each
x∗ ∈ F, we obtain
G x∗ , Jyn Gx∗ , αn Jzn 1 − αn Jxn x∗ 2 − 2αn x∗ , Jzn − 21 − αn x∗ , Jxn αn Jzn 1 − αn Jxn 2 2ρfx∗ ≤ x∗ 2 − 2αn x∗ , Jzn − 21 − αn x∗ , Jxn αn Jzn 2 1 − αn Jxn 2 2ρfx∗ αn Gx∗ , Jzn 1 − αn Gx∗ , Jxn 1
2
3
1 − αn Gx∗ , Jxn αn G x∗ , βn Jx0 βn JTn xn βn JSn xn
≤ 1 − αn Gx∗ , Jxn Abstract and Applied Analysis
1
2
3
αn x∗ 2 − 2βn x∗ , Jx0 − 2βn x∗ , JTn xn − 2βn x∗ , JSn xn 1
2
3
βn x0 2 βn Tn xn 2 βn Sn xn 2 2ρfx∗ 9
1
2
3
1 − αn Gx∗ , Jxn αn βn Gx∗ , Jx0 βn Gx∗ , JTn xn βn Gx∗ , JSn xn 1
2
3
≤ 1 − αn Gx∗ , Jxn αn βn Gx∗ , Jx0 βn Gx∗ , Jxn βn Gx∗ , Jxn 1
1
≤ 1 − αn βn Gx∗ , Jxn αn βn Gx∗ , Jx0 .
3.3
So, x∗ ∈ Cn . This implies that F ⊂ Cn , for all n ≥ 0. Since Cn1 is closed and convex for all n ≥
f
0 and ∅ /
F ⊂ Cn , for all n ≥ 0, it follows that ΠCn1 x0 is well defined for all n ≥ 0.
We now show that limn → ∞ Gxn , Jx0 exists. Since f : E → R is a convex and lower
semicontinuous, applying Lemma 2.5, we see that there exists u∗ ∈ E∗ and α ∈ R such that
f y ≥ y, u∗ α,
∀y ∈ E.
3.4
It follows that
Gxn , Jx0 xn 2 − 2xn , Jx0 x0 2 2ρfxn ≥ xn 2 − 2xn , Jx0 x0 2 2ρxn , u∗ 2ρα
xn 2 − 2 xn , Jx0 − ρu∗ x0 2 2ρα
≥ xn 2 − 2xn Jx0 − ρu∗ x0 2 2ρα
2
2
xn − Jx0 − ρu∗ x0 2 − Jx0 − ρu∗ 2ρα.
3.5
f
Since xn ΠCn x0 , it follows from 3.5 that
2
2
Gx∗ , Jx0 ≥ Gxn , Jx0 ≥ xn − Jx0 − ρu∗ x0 2 − Jx0 − ρu∗ 2ρα
3.6
∞
for each x∗ ∈ F. This implies that {xn }∞
n0 is bounded and so is {Gxn , Jx0 }n0 . By the
f
construction of Cn , we have that Cm ⊂ Cn and xm ΠCm x0 ∈ Cn for any positive integer
m ≥ n. It then follows Lemma 2.7 that
φxm , xn Gxn , Jx0 ≤ Gxm , Jx0 .
3.7
φxm , xn ≥ xm − xn 2 ≥ 0.
3.8
It is obvious that
10
Abstract and Applied Analysis
In particular,
φxn1 , xn Gxn , Jx0 ≤ Gxn1 , Jx0 ,
3.9
φxn1 , xn ≥ xn1 − xn 2 ≥ 0,
∞
and so {Gxn , Jx0 }∞
n0 is nondecreasing. It follows that the limit of {Gxn , Jx0 }n0 exists.
Now, 3.7 implies that
φxm , xn ≤ Gxm , Jx0 − Gxn , Jx0 .
3.10
Taking the limit as m, n → ∞ in 3.10, we obtain
lim φxm , xn 0.
3.11
n→∞
It then follows from Lemma 2.9 that xm − xn → 0 as m, n → ∞. Hence, {xn }∞
n0 is Cauchy.
Since E is a Banach space and C is closed and convex, then there exists p ∈ C such that
xn → p as n → ∞.
Now since φxm , xn → 0 as m, n → ∞ we have in particular that φxn1 , xn →
f
0 as n → ∞ and this further implies that limn → ∞ xn1 − xn 0. Since xn1 ΠCn1 x0 ∈ Cn1 ,
we have
1
1
φ xn1 , yn ≤ 1 − αn βn φxn1 , xn αn βn φxn1 , x0 ,
∀n ≥ 0.
3.12
Then, we obtain
lim φ xn1 , yn 0.
n→∞
3.13
Since E is uniformly convex and smooth, we have from Lemma 2.9 that
lim xn1 − xn 0 lim xn1 − yn .
3.14
xn − yn ≤ xn1 − xn xn1 − yn .
3.15
lim xn − yn 0.
3.16
n→∞
n→∞
So,
Hence,
n→∞
Since J is uniformly norm-to-norm continuous on bounded sets and limn → ∞ xn − yn 0,
we obtain
lim Jxn − Jyn 0.
n→∞
3.17
Abstract and Applied Analysis
11
Since 0 < α ≤ αn < 1, then
Jzn − Jxn 1
Jxn − Jyn −→ 0,
αn
as n → ∞.
3.18
Since J −1 is uniformly norm-to-norm continuous on bounded sets, we have
lim zn − xn 0,
3.19
n→∞
so that zn → p as n → ∞. Since {xn } is bounded, so are {zn }, {JTn xn }, and {JSn xn }.
Let r : supn≥0 {xn , Tn xn , Sn xn }. Then from Lemma 2.10, we have
G x∗ , Jyn Gx∗ , αn Jzn 1 − αn Jxn x∗ 2 − 2αn x∗ , Jzn − 21 − αn x∗ , Jxn αn Jzn 1 − αn Jxn 2 2ρfx∗ ≤ x∗ 2 − 2αn x∗ , Jzn − 21 − αn x∗ , Jxn αn Jzn 2 1 − αn Jxn 2 2ρfx∗ αn Gx∗ , Jzn 1 − αn Gx∗ , Jxn 1
2
3
1 − αn Gx∗ , Jxn αn G x∗ , βn Jx0 βn JTn xn βn JSn xn
≤ 1 − αn Gx∗ , Jxn 1
2
3
1
αn x∗ 2 − 2βn x∗ , Jx0 − 2βn x∗ , JTn xn − 2βn x∗ , JSn xn βn x0 2
2
3
2 3
βn Tn xn 2 βn Sn xn 2 − βn βn gJTn xn − JSn xn 2ρfx∗ 1
2
3
1 − αn Gx∗ , Jxn αn βn Gx∗ , Jx0 βn Gx∗ , JTn xn βn Gx∗ , JSn xn 2 3
− βn βn gJTn xn − JSn xn 1
2
3
≤ 1 − αn Gx∗ , Jxn αn βn Gx∗ , Jx0 βn Gx∗ , Jxn βn Gx∗ , Jxn 2 3
− βn βn gJTn xn − JSn xn 1
≤ 1 − αn Gx∗ , Jxn βn Gx∗ , Jx0 αn Gx∗ , Jxn 2 3
− αn βn βn gJTn xn − JSn xn 1
2 3
βn Gx∗ , Jx0 Gx∗ , Jxn − αn βn βn gJTn xn − JSn xn .
3.20
It then follows that
2 3
1
αn βn βn gJTn xn − JSn xn ≤ βn Gx∗ , Jx0 Gx∗ , Jxn − G x∗ , Jyn .
3.21
12
Abstract and Applied Analysis
But
2
Gx∗ , Jxn − G x∗ , Jyn xn 2 − yn − 2 x∗ , Jxn − Jyn
2 ≤ xn 2 − yn 2 x∗ , Jxn − Jyn ≤ xn − yn xn yn 2x∗ Jxn − Jyn ≤ xn − yn xn yn 2x∗ Jxn − Jyn .
3.22
From limn → ∞ xn − yn 0 and limn → ∞ Jxn − Jyn 0, we obtain
Gx∗ , Jxn − G x∗ , Jyn −→ 0,
n → ∞.
3.23
2 3
Using the condition lim infn → ∞ βn βn > 0 and αn ≥ α > 0, we have
lim gJTn xn − JSn xn 0.
n→∞
3.24
By property of g, we have limn → ∞ JTn xn − JSn xn 0. Since J −1 is also uniformly norm-tonorm continuous on bounded sets, we have
lim Tn xn − Sn xn 0.
n→∞
3.25
Furthermore,
1
2
3
Jxn − Jzn Jxn − βn Jx0 βn JTn xn βn JSn xn 1
2
3
βn Jxn − Jx0 βn Jxn − JTn xn βn Jxn − JSn xn 2
1
3
≥ βn Jxn − JTn xn βn Jxn − JSn xn − βn Jxn − Jx0 .
3.26
This implies that
2
3
1
βn Jxn − JTn xn βn Jxn − JSn xn ≤ Jxn − Jzn βn Jxn − Jx0 .
3.27
1
Since xn → p, zn → p and limn → ∞ βn 0, it follows from 3.27 that
2
3
lim βn Jxn − JTn xn βn Jxn − JSn xn 0.
n→∞
3.28
Abstract and Applied Analysis
13
On the other hand, by using the property of norm, we obtain
2
3
βn Jxn − JTn xn βn Jxn − JSn xn 2
3
βn Jxn − JTn xn βn Jxn − JSn xn 3
3
βn Jxn − JTn xn − βn Jxn − JTn xn 2
3
3
βn βn Jxn − JTn xn βn JTn xn − JSn xn 3.29
2
3
3
≥ βn βn Jxn − JTn xn − βn JTn xn − JSn xn .
This follows that
2
2
3
3
3
βn βn Jxn − JTn xn ≤ βn Jxn − JTn xn βn Jxn − JSn xn βn JTn xn − JSn xn .
3.30
By 3.27 and limn → ∞ JTn xn − JSn xn 0, we have
2
3
lim βn βn Jxn − JTn xn 0.
3.31
n→∞
2
3
Since 0 < b ≤ βn ≤ 1 and 0 < d ≤ βn ≤ 1, we have
lim xn − Tn xn 0.
3.32
lim xn − Sn xn 0.
3.33
n→∞
Similarly, we can show that
n→∞
∞
Since xn → p and {Tn }, {Sn } are uniformly closed, we have p ∈ ∞
n0 FTn ∩ n0 FSn .
∞
f
Finally, we show that p ΠF x0 . Since F ∞
n0 FTn ∩ n0 FSn is a closed and
f
f
convex set, from Lemma 2.6, we know that ΠF x0 is single valued and denote w ΠF x0 . Since
f
xn ΠCn x0 and w ∈ F ⊂ Cn , we have
Gxn , Jx0 ≤ Gw, Jx0 ,
∀n ≥ 0.
3.34
We know that Gξ, Jϕ is convex and lower semicontinuous with respect to ξ when ϕ is fixed.
This implies that
G p, Jx0 ≤ lim inf Gxn , Jx0 ≤ lim sup Gxn , Jx0 ≤ Gw, Jx0 .
n→∞
f
n→∞
From the definition of ΠF x0 and p ∈ F, we see that p w. This completes the proof.
3.35
14
Abstract and Applied Analysis
Corollary 3.2. Let E be a uniformly convex real Banach space which is also uniformly smooth. Let
∞
two countable
C be a nonempty, closed, and convex subset of E. Suppose {Tn }∞
n0 and {Sn }n0 are families of weak relatively nonexpansive mappings of C into itself such that F : ∞
n0 FTn ∩
∞
FS
∅.
Suppose
{x
}
is
iteratively
generated
by
x
∈
C,
C
C
∞
n /
n n0
0
0
n0
1
2
3
zn J −1 βn Jx0 βn JTn xn βn JSn xn ,
Cn1
yn J −1 αn Jzn 1 − αn Jxn ,
1
1
w ∈ Cn : φ w, yn ≤ 1 − αn βn φw, xn αn βn φw, x0 ,
xn1 ΠCn1 x0 ,
3.36
n ≥ 0,
with the conditions
1
i limn → ∞ βn 0;
2
3
ii 0 < b ≤ βn ≤ 1, 0 < d ≤ βn ≤ 1;
iii 0 < α ≤ αn ≤ 1 for some α ∈ 0, 1.
Then, {xn }∞
n0 converges strongly to ΠF x0 .
f
Proof. Take fx 0 for all x ∈ E in Theorem 3.1, Gξ, Jx φξ, x and ΠC x0 ΠC x0 . Then,
the desired conclusion follows.
Remark 3.3. Our Corollary 3.2 extends and improves on Theorem 1.5. In fact, the iterative
procedure 3.36 is simpler than 1.8 in the following two aspects: a the process of
computing Qn {w ∈ Cn−1 ∩ Qn−1 : xn − w, Jx0 − Jxn ≥ 0} is removed; b the process
of computing ΠCn ∩Qn is replaced by computing ΠCn .
4. Applications
A mapping H from E to E∗ is said to be
i monotone if Hx − Hy, x − y ≥ 0, for all x, y ∈ E;
ii strictly monotone if H is monotone and Hx − Hy, x − y 0 if and only if x y;
iii β-Lipschitz continuous if there exists a constant β ≥ 0 such that Hx − Hy ≤
βx − y, for all x, y ∈ E.
Let M be a set valued mapping from E to E∗ with domain DM {z ∈ E : Mz /
∅}
and range RM {Mz : z ∈ DM}. A set-valued mapping M is said to be
i monotone if x1 − x2 , y1 − y2 ≥ 0 for each xi ∈ DM and yi ∈ Mxi , i 1, 2,
ii r-strongly monotone if x1 − x2 , y1 − y2 ≥ rx1 − x2 2 for each xi ∈ DM and
yi ∈ Mxi , i 1, 2,
iii maximal monotone if M is monotone and its graph GM : {x, y : y ∈ Mx} is
not properly contained in the graph of any other monotone operator,
iv a general H-monotone if M is monotone and H λME E∗ holds for every
λ > 0, where H is a mapping from E to E∗ .
Abstract and Applied Analysis
15
We denote the set {x ∈ E : 0 ∈ Mx} by M−1 0. From Li et al. 15, we know that
∗
if H : E → E∗ is strictly monotone and M : E → 2E is general H-monotone mapping,
−1
then M 0 is closed and convex. Furthermore, for every λ > 0 and x∗ ∈ E∗ , there exists a
unique x ∈ DM such that x H λM−1 x∗ . Thus, we can define a single-value mapping
Tλ : E → DM by Tλ x H λM−1 Hx. It is obvious that M−1 0 FTλ for all λ > 0.
Lemma 4.1 Alber, 16. If E is a uniformly convex and uniformly smooth Banach space, δE is
the modulus of convexity of E and ρE t is the modulus of smoothness of E, then the inequalities
2
8d δE
x − ξ
4d
hold for all x and ξ in E, where d ≤ φx, ξ ≤ 4d ρE
2
4x − ξ
d
4.1
x2 ξ2 /2.
Lemma 4.2 Xia and Huang, 22. Let E be a Banach space with dual space E∗ , H : E → E∗ a
∗
strictly monotone mapping and M : E → 2E a general H-monotone mapping. Then
i H λM−1 is a single valued mapping;
ii if E is reflexive and M : E → 2E is r-strongly monotone, H λM−1 is Lipschitz
continuous with constant 1/λr, where r > 0.
∗
Theorem 4.3. Let E be a uniformly convex real Banach space which is also uniformly smooth with
δE ≥ k2 and ρE t ≤ ct2 for some k, c > 0. Suppose H : E → E∗ is a strictly monotone
∗
and β-Lipschitz continuous mapping and Mi : E → 2E is a general H-monotone mapping
and ri -strongly monotone mapping with ri > 0, i 1, 2 such that F : M1−1 0 ∩ M2−1 0 / ∅.
Mi
−1
Let Tλ H λMi H, i 1, 2 and f : E → R be a convex and lower semicontinuous
mapping with Df E and suppose for each n ≥ 0, there exists a λn > 0 such that 64cβ2 ≤
min{1/2kλ2n r12 , 1/2kλ2n r12 }. Let {xn }∞
n0 be iteratively generated by x0 ∈ E, C0 E,
1
2
3
zn J −1 βn Jx0 βn JTλMn 1 xn βn JTλMn 2 xn ,
Cn1
yn J −1 αn Jxn 1 − αn Jzn ,
1
1
w ∈ Cn : G w, Jyn ≤ 1 − αn βn Gw, Jxn αn βn Gw, Jx0 ,
f
xn1 ΠCn1 x0 ,
with the conditions
1
i limn → ∞ βn 0;
2
3
ii 0 < b ≤ βn ≤ 1, 0 < d ≤ βn ≤ 1;
iii 0 < α ≤ αn ≤ 1 for some α ∈ 0, 1;
iv lim infn → ∞ λn > 0.
Then, {xn }∞
n0 converges strongly to ΠF x0 .
f
n ≥ 0,
4.2
16
Abstract and Applied Analysis
Proof. We only need to prove that {TλMn 1 } and {TλMn 2 } are countable families of weak
M1
−1
relatively nonexpansive mappings with common fixed points sets ∞
n0 FTλn M1 0 and
∞
M2
M
∞
−1
−1
1
∅. Secondly, we
n0 FTλn M2 0, respectively. Firstly, we have
n0 FTλn M1 0 /
show that φp, TλMn 1 w ≤ φp, w, for all w ∈ E, p ∈ FTλMn 1 , n ≥ 0. Now, by Lemma 4.2 and
the Lipschitz continuity of H, we have
M1
M
Tλn p − Tλn 1 w H λn M1 −1 Hp − H λn M1 −1 Hw
≤
1 Hp − Hw
λn r1
≤
β p − w.
λn r1
4.3
By 4.3 and Lemma 4.1,
φ p, TλMn 1 w φ TλMn 1 p, TλMn 1 w
⎞
⎛ 4TλMn 1 p − TλMn 1 w
2
⎝
⎠
≤ 4d ρE
d
2
≤ 64cTλMn 1 p − TλMn 1 w
4.4
64cβ2 p − w2 ,
2 2
λn r1
p − w
2
1 2
≥ kp − w .
φ p, w ≥ 8d δE
4d
2
≤
Since 64cβ2 ≤ 1/2kλ2n r12 , it follows from 4.4 that φp, TλMn 1 w ≤ φp, w, for all w ∈ E, p ∈
M1
M1
} ∞
M−1 0. We first show that
FT M1 , n ≥ 0. Thirdly, we show that F{T
n0 FT
λn
λn
λn
1
M1
M1
} ⊂ M1−1 0. Let p ∈ F{T
}, then there exists {xn } ⊂ E such that xn → p and
F{T
λn
λn
limn → ∞ xn − TλMn 1 xn 0. Since H is β-Lipschitz continuous,
M
M
Hxn − HTλn 1 xn ≤ βxn − Tλn 1 xn .
4.5
1
Hxn − HTλMn 1 xn −→ 0.
λn
4.6
Letting n → ∞, we obtain
Abstract and Applied Analysis
17
It follows from 1/λn Hxn − HTλMn 1 xn ∈ M1 TλMn 1 xn and the monotonicity of M1 that
x − TλMn 1 xn , x∗ −
1
Hxn − HTλMn 1 xn
λn
≥0
4.7
for all x ∈ DM1 and x∗ ∈ M1 x. Taking the limit as n → ∞, we obtain
x − p, x∗ ≥ 0
4.8
for all x ∈ DM1 and x∗ ∈ M1 x. By the maximality of M1 , we know that p ∈ M1−1 0. On
M1 for all n ≥ 0, therefore,
the other hand, we know that FTλMn 1 M1−1 0, FTλMn 1 ⊂ FT
λn
M
M
∞
∞
n0 T 1 . Thus, we have proved that {T M1 } is a countable family
M1−1 0 n0 FTλn 1 F
λn
λn
M1
of weak relatively nonexpansive mappings with common fixed points sets ∞
n0 FTλn M1−1 0. By following the same arguments, we can show that {TλMn 2 } is a countable family of
M2
−1
weak relatively nonexpansive mappings with common fixed points sets ∞
n0 FTλn M2 0.
Let E be a uniformly convex and uniformly smooth Banach space, H J and M
be a maximal monotone mapping. Then, we can define Jλ J λM−1 J for all λ > 0. We
know that Jλ is relatively nonexpansive and therefore weak relatively nonexpansive and
M−1 0 FJλ for all λ > 0 see, e.g., 2, where FJλ denotes the fixed points set of Jλ .
By Corollary 3.2, we obtain the following theorem.
Theorem 4.4. Let E be a uniformly convex real Banach space which is also uniformly smooth. For
each i 1, 2, let Mi ⊂ E×E∗ be a maximal monotone operator and let JλMi J λMi −1 J for all λ > 0
and suppose C is a nonempty closed and convex subset of E such that DMi ⊂ C ⊂ J −1 λ>0 RJ ∞
λMi and F : M1−1 0 M2−1 0 / ∅. Let {xn }n0 be iteratively generated by x0 ∈ E, C0 E,
1
2
3
zn J −1 βn Jx0 βn JJTλMn 1 xn βn JJTλMn 2 xn ,
Cn1
yn J −1 αn Jxn 1 − αn Jzn ,
1
1
w ∈ Cn : φ w, yn ≤ 1 − αn βn φw, xn αn βn φw, x0 ,
xn1 ΠCn1 x0 ,
with the conditions
1
i limn → ∞ βn 0;
2
3
ii 0 < b ≤ βn ≤ 1, 0 < d ≤ βn ≤ 1;
iii 0 < α ≤ αn ≤ 1 for some α ∈ 0, 1;
iv lim infn → ∞ λn > 0.
Then, {xn }∞
n0 converges strongly to ΠF x0 .
n ≥ 0,
4.9
18
Abstract and Applied Analysis
Acknowledgment
The author would like to express his thanks to Professor Jean Pierre Gossez and the referees
for their valuable suggestions and comments.
References
1 C. Chidume, Geometric Properties of Banach Spaces and Nonlinear Iterations, vol. 1965 of Lecture Notes in
Mathematics, Springer, London, UK, 2009.
2 W. Takahashi, Nonlinear Functional Analysis-Fixed Point Theory and Application, Yokohama Publishers,
Yokohama, Japan, 2000.
3 W. Takahashi, Nonlinear Functional Analysis, Yokohama Publishers, Yokohama, Japan, 2000.
4 S. Kamimura and W. Takahashi, “Strong convergence of a proximal-type algorithm in a Banach
space,” SIAM Journal on Optimization, vol. 13, no. 3, pp. 938–945, 2002.
5 Y. Su, H.-K. Xu, and X. Zhang, “Strong convergence theorems for two countable families of
weak relatively nonexpansive mappings and applications,” Nonlinear Analysis. Theory, Methods &
Applications, vol. 73, no. 12, pp. 3890–3906, 2010.
6 D. Butnariu, S. Reich, and A. J. Zaslavski, “Asymptotic behavior of relatively nonexpansive operators
in Banach spaces,” Journal of Applied Analysis, vol. 7, no. 2, pp. 151–174, 2001.
7 D. Butnariu, S. Reich, and A. J. Zaslavski, “Weak convergence of orbits of nonlinear operators in
reflexive Banach spaces,” Numerical Functional Analysis and Optimization, vol. 24, no. 5-6, pp. 489–508,
2003.
8 Y. Censor and S. Reich, “Iterations of paracontractions and firmly nonexpansive operators with
applications to feasibility and optimization,” Optimization, vol. 37, no. 4, pp. 323–339, 1996.
9 S.-Y. Matsushita and W. Takahashi, “A strong convergence theorem for relatively nonexpansive
mappings in a Banach space,” Journal of Approximation Theory, vol. 134, no. 2, pp. 257–266, 2005.
10 X. Qin and Y. Su, “Strong convergence theorems for relatively nonexpansive mappings in a Banach
space,” Nonlinear Analysis. Theory, Methods & Applications, vol. 67, no. 6, pp. 1958–1965, 2007.
11 W. Takahashi and K. Zembayashi, “Strong convergence theorem by a new hybrid method for
equilibrium problems and relatively nonexpansive mappings,” Fixed Point Theory and Applications,
vol. 2008, Article ID 528476, 11 pages, 2008.
12 J. Kang, Y. Su, and X. Zhang, “Hybrid algorithm for fixed points of weak relatively nonexpansive
mappings and applications,” Nonlinear Analysis. Hybrid Systems, vol. 4, no. 4, pp. 755–765, 2010.
13 H. Zegeye and N. Shahzad, “Strong convergence theorems for a finite family of asymptotically
nonexpansive mappings and semigroups,” Nonlinear Analysis. Theory, Methods & Applications, vol.
69, no. 12, pp. 4496–4503, 2008.
14 S. Plubtieng and K. Ungchittrakool, “Strong convergence theorems for a common fixed point of two
relatively nonexpansive mappings in a Banach space,” Journal of Approximation Theory, vol. 149, no. 2,
pp. 103–115, 2007.
15 X. Li, N.-J. Huang, and D. O’Regan, “Strong convergence theorems for relatively nonexpansive
mappings in Banach spaces with applications,” Computers & Mathematics with Applications, vol. 60,
no. 5, pp. 1322–1331, 2010.
16 Y. I. Alber, “Metric and generalized projection operators in Banach spaces: properties and
applications,” in Theory and Applications of Nonlinear Operators of Accretive and Monotone Type, vol.
178 of Lecture Notes in Pure and Applied Mathematics, pp. 15–50, Marcel Dekker, New York, NY, USA,
1996.
17 Y. I. Alber and S. Reich, “An iterative method for solving a class of nonlinear operator equations in
Banach spaces,” Panamerican Mathematical Journal, vol. 4, no. 2, pp. 39–54, 1994.
18 I. Cioranescu, Geometry of Banach Spaces, Duality Mappings and Nonlinear Problems, vol. 62 of
Mathematics and its Applications, Kluwer Academic, Dordrecht, The Netherlands, 1990.
19 K.-Q. Wu and N.-J. Huang, “The generalised f-projection operator with an application,” Bulletin of
the Australian Mathematical Society, vol. 73, no. 2, pp. 307–317, 2006.
20 J. Fan, X. Liu, and J. Li, “Iterative schemes for approximating solutions of generalized variational
inequalities in Banach spaces,” Nonlinear Analysis. Theory, Methods & Applications, vol. 70, no. 11, pp.
3997–4007, 2009.
Abstract and Applied Analysis
19
21 Y. J. Cho, H. Zhou, and G. Guo, “Weak and strong convergence theorems for three-step iterations
with errors for asymptotically nonexpansive mappings,” Computers & Mathematics with Applications,
vol. 47, no. 4-5, pp. 707–717, 2004.
22 F.-Q. Xia and N.-J. Huang, “Variational inclusions with a general H-monotone operator in Banach
spaces,” Computers & Mathematics with Applications, vol. 54, no. 1, pp. 24–30, 2007.
Advances in
Operations Research
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Advances in
Decision Sciences
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Mathematical Problems
in Engineering
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Journal of
Algebra
Hindawi Publishing Corporation
http://www.hindawi.com
Probability and Statistics
Volume 2014
The Scientific
World Journal
Hindawi Publishing Corporation
http://www.hindawi.com
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
International Journal of
Differential Equations
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Volume 2014
Submit your manuscripts at
http://www.hindawi.com
International Journal of
Advances in
Combinatorics
Hindawi Publishing Corporation
http://www.hindawi.com
Mathematical Physics
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Journal of
Complex Analysis
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
International
Journal of
Mathematics and
Mathematical
Sciences
Journal of
Hindawi Publishing Corporation
http://www.hindawi.com
Stochastic Analysis
Abstract and
Applied Analysis
Hindawi Publishing Corporation
http://www.hindawi.com
Hindawi Publishing Corporation
http://www.hindawi.com
International Journal of
Mathematics
Volume 2014
Volume 2014
Discrete Dynamics in
Nature and Society
Volume 2014
Volume 2014
Journal of
Journal of
Discrete Mathematics
Journal of
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Applied Mathematics
Journal of
Function Spaces
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Optimization
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Hindawi Publishing Corporation
http://www.hindawi.com
Volume 2014
Download